Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network
Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. We use genetic algorithm (GA) optimization to discover unconventional spin-crossover comp...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
American Chemical Society (ACS)
2019
|
| Online Access: | http://hdl.handle.net/1721.1/120162 https://orcid.org/0000-0001-7825-4797 https://orcid.org/0000-0001-9342-0191 |
| _version_ | 1826210234445594624 |
|---|---|
| author | Janet, Jon Paul Chan, Lydia C. Kulik, Heather Janine |
| author2 | Massachusetts Institute of Technology. Department of Chemical Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Chemical Engineering Janet, Jon Paul Chan, Lydia C. Kulik, Heather Janine |
| author_sort | Janet, Jon Paul |
| collection | MIT |
| description | Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. We use genetic algorithm (GA) optimization to discover unconventional spin-crossover complexes in combination with efficient scoring from an artificial neural network (ANN) that predicts spin-state splitting of inorganic complexes. We explore a compound space of over 5600 candidate materials derived from eight metal/oxidation state combinations and a 32-ligand pool. We introduce a strategy for error-aware ML-driven discovery by limiting how far the GA travels away from the nearest ANN training points while maximizing property (i.e., spin-splitting) fitness, leading to discovery of 80% of the leads from full chemical space enumeration. Over a 51-complex subset, average unsigned errors (4.5 kcal/mol) are close to the ANN's baseline 3 kcal/mol error. By obtaining leads from the trained ANN within seconds rather than days from a DFT-driven GA, this strategy demonstrates the power of ML for accelerating inorganic material discovery. |
| first_indexed | 2024-09-23T14:46:17Z |
| format | Article |
| id | mit-1721.1/120162 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T14:46:17Z |
| publishDate | 2019 |
| publisher | American Chemical Society (ACS) |
| record_format | dspace |
| spelling | mit-1721.1/1201622022-09-29T10:27:03Z Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network Janet, Jon Paul Chan, Lydia C. Kulik, Heather Janine Massachusetts Institute of Technology. Department of Chemical Engineering Janet, Jon Paul Chan, Lydia C. Kulik, Heather Janine Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. We use genetic algorithm (GA) optimization to discover unconventional spin-crossover complexes in combination with efficient scoring from an artificial neural network (ANN) that predicts spin-state splitting of inorganic complexes. We explore a compound space of over 5600 candidate materials derived from eight metal/oxidation state combinations and a 32-ligand pool. We introduce a strategy for error-aware ML-driven discovery by limiting how far the GA travels away from the nearest ANN training points while maximizing property (i.e., spin-splitting) fitness, leading to discovery of 80% of the leads from full chemical space enumeration. Over a 51-complex subset, average unsigned errors (4.5 kcal/mol) are close to the ANN's baseline 3 kcal/mol error. By obtaining leads from the trained ANN within seconds rather than days from a DFT-driven GA, this strategy demonstrates the power of ML for accelerating inorganic material discovery. United States. Office of Naval Research (Grant N00014-17-1-2956) United States. Department of Energy (Grant DE-SC0018096) National Science Foundation (U.S.) (Grant CBET-1704266) 2019-02-04T14:37:34Z 2019-02-04T14:37:34Z 2018-02 2018-01 2019-02-01T13:23:51Z Article http://purl.org/eprint/type/JournalArticle 1948-7185 http://hdl.handle.net/1721.1/120162 Janet, Jon Paul et al. “Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network.” The Journal of Physical Chemistry Letters 9, 5 (February 2018): 1064–1071 © 2018 American Chemical Society https://orcid.org/0000-0001-7825-4797 https://orcid.org/0000-0001-9342-0191 http://dx.doi.org/10.1021/ACS.JPCLETT.8B00170 Journal of Physical Chemistry Letters Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf American Chemical Society (ACS) ACS |
| spellingShingle | Janet, Jon Paul Chan, Lydia C. Kulik, Heather Janine Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title | Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title_full | Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title_fullStr | Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title_full_unstemmed | Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title_short | Accelerating Chemical Discovery with Machine Learning: Simulated Evolution of Spin Crossover Complexes with an Artificial Neural Network |
| title_sort | accelerating chemical discovery with machine learning simulated evolution of spin crossover complexes with an artificial neural network |
| url | http://hdl.handle.net/1721.1/120162 https://orcid.org/0000-0001-7825-4797 https://orcid.org/0000-0001-9342-0191 |
| work_keys_str_mv | AT janetjonpaul acceleratingchemicaldiscoverywithmachinelearningsimulatedevolutionofspincrossovercomplexeswithanartificialneuralnetwork AT chanlydiac acceleratingchemicaldiscoverywithmachinelearningsimulatedevolutionofspincrossovercomplexeswithanartificialneuralnetwork AT kulikheatherjanine acceleratingchemicaldiscoverywithmachinelearningsimulatedevolutionofspincrossovercomplexeswithanartificialneuralnetwork |